U.S. patent application number 12/532774 was filed with the patent office on 2010-08-26 for angiogenesis monitoring using in vivo hyperspectral radiometric imaging.
Invention is credited to Wafik S. El-Deiry, Jeremy M. Lerner, Paul Tumeh.
Application Number | 20100217129 12/532774 |
Document ID | / |
Family ID | 40226707 |
Filed Date | 2010-08-26 |
United States Patent
Application |
20100217129 |
Kind Code |
A1 |
El-Deiry; Wafik S. ; et
al. |
August 26, 2010 |
ANGIOGENESIS MONITORING USING IN VIVO HYPERSPECTRAL RADIOMETRIC
IMAGING
Abstract
This invention relates to the use of in-vivo hyperspectral
imaging to monitor angiogenesis. Specifically, the invention
provides systems and methods of obtaining hyperspectral images of a
field of view comprising an area sought to be monitored.
Inventors: |
El-Deiry; Wafik S.; (Bryn
Mawr, PA) ; Tumeh; Paul; (Philadelphia, PA) ;
Lerner; Jeremy M.; (Hillsborough, NJ) |
Correspondence
Address: |
Pearl Cohen Zedek Latzer, LLP
1500 Broadway, 12th Floor
New York
NY
10036
US
|
Family ID: |
40226707 |
Appl. No.: |
12/532774 |
Filed: |
March 20, 2008 |
PCT Filed: |
March 20, 2008 |
PCT NO: |
PCT/US08/03678 |
371 Date: |
April 19, 2010 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
60907165 |
Mar 23, 2007 |
|
|
|
Current U.S.
Class: |
600/476 |
Current CPC
Class: |
A61B 5/0059 20130101;
A61B 5/418 20130101; A61B 5/415 20130101; A61B 2562/223 20130101;
A61B 5/489 20130101 |
Class at
Publication: |
600/476 |
International
Class: |
A61B 6/00 20060101
A61B006/00 |
Goverment Interests
GOVERNMENT INTEREST
[0001] This invention was supported, in part, by Grant Number T32
NS043126-03 from the NIH; and Grant Number U54105008 from the
National Cancer Institute's Network for Translational Research in
Optical Imaging (NTROI). The government may have certain rights in
the invention.
Claims
1. A hyperspectral imaging system comprising: an electromagnetic
energy source; coupled to a prism and reflector imaging
spectroscopy system (PARISS) equipped with an imaging probe
comprising a bundle of structured optical fibers capable of being
used for a spatially resolved imaging.
2. The system of claim 1, wherein the electromagnetic energy source
is a high intensity light source.
3. The system of claim 2, wherein the high intensity light source
is a high intensity tungsten halogen lamp, or a Xenon lamp.
4. The system of claim 1, wherein the plurality of customized
bundle of structured optical fibers that can be used as spatially
resolved imaging probes are spatially discrete.
5. The system of claim 1, wherein the prism and reflector imaging
spectroscopy system comprises an imaging spectrometer integrated
with a first camera; and a second camera acting as an observed
image camera.
6. The system of claim 5, wherein the first and second camera is a
QICAM.
7. The system of claim 1, wherein the plurality of customized
bundle of structured optical fibers that can be used as spatially
resolved imaging probes comprise illumination probes and signal
collection probes.
8. The system of claim 7, wherein the system collection probes are
arrayed along a slit.
9. The system of claim 8, wherein the system collection probes
collect images onto an entrance slit in the prism and reflector
imaging spectroscopy system.
10. The system of claim 7, wherein the illumination probes and
signal collection probes comprise no less than 16 illumination
fibers and no less than 16 signal collection fibers.
11. The system of claim 1, wherein the remote fiber-optic probes
are capable of being mapped to produce an image.
12. The system of claim 1, wherein the remote fiber-optic probes
are distributed in a silicone cuff to be placed over an organ, a
tissue or their combination.
13. A method of acquiring in-vivo hyperspectral image from a
subject, comprising: selecting a field of view (FOV) of the
subject; attaching a plurality of customized remote fiber-optic
probes, wherein the customized remote fiber-optic probes are
operably linked to a hyperspectral imaging system comprising: an
electromagnetic energy source; coupled to a prism and reflector
imaging spectroscopy system (PARISS) equipped with an imaging
probe; illuminating the field of view using the hyperspectral
imaging system; and collecting an in-vivo hyperspectral image.
14. The method of claim 13, whereby wherein the electromagnetic
energy source is a high intensity light source.
15. The method of claim 14, whereby the high intensity light source
is a high intensity tungsten halogen lamp.
16. The method of claim 13, whereby the plurality of customized
bundle of structured optical fibers that can be used as a spatially
resolved imaging probes are spatially discrete.
17. The method of claim 13, whereby the prism and reflector imaging
spectroscopy system comprises an imaging spectrometer integrated
with a first camera; and a second camera acting as an observed
image camera.
18. The method of claim 17, whereby the first or second camera or
both are QICAM
19. The method of claim 13, whereby the plurality of customized
bundle of structured optical fibers that can be used as spatially
resolved imaging probes comprise illumination probes and signal
collection probes.
20. The method of claim 19, whereby the collection probes are
arrayed along a slit.
21. The method of claim 20, whereby the collection probes collect
images onto an entrance slit in the a prism and reflector imaging
spectroscopy system.
22. The method of claim 19, whereby the illumination probes and
signal collection probes comprise no less than 15 illumination
fibers and no less than 16 signal collection fibers.
23. The method of claim 13, further comprising compiling a unique
spectral signature of the field of view.
24. An image acquired by the method of claim 13.
25. The spectral signature of a field of view (FOV) compiled by the
method of claim 23.
26. A method of monitoring neoplasia of a tissue in a subject,
comprising the step of obtaining a hyperspectral image, according
to the method of claim 13, of a field of view of an area sought to
be monitored; and comparing the image to a standard.
27-37. (canceled)
38. The method of claim 26, further comprising comparing the
spectral signature of the field of view with the spectral signature
of the same field of view obtained from a subject exhibiting
neoplasia.
39. The method of claim 26, further comprising comparing the
spectral signature of the field of view with the spectral signature
of the same field of view obtained from a subject not exhibiting
neoplasia.
40. The method of claim 26, whereby the standard is a hyperspectral
image of the tissue at a predetermined point.
41. The method of claim 40, whereby the predetermined point is
time, course of treatment, dosage of a therapeutic agent or their
combination.
42. A method of imaging a natural history or response to therapy of
lesions of the skin, oropharynx, esophagus, bladder, or
intra-abdominal lesions accessed through laparoscopy, comprising
the step of obtaining a hyperspectral image, according to the
method of claim 13, of a field of view of an area sought to be
monitored in the lesions of the skin, oropharynx, esophagus,
bladder, or intra-abdominal lesions accessed through laparoscopy;
and comparing the image to a standard.
43-53. (canceled)
54. The method of claim 43, further comprising comparing the
spectral signature of the field of view with the spectral signature
of the same field of view obtained from a healthy subject.
55. The method of claim 42, whereby the standard is a hyperspectral
image of the tissue at a predetermined point.
56. The method of claim 55, whereby the predetermined point is
time, course of treatment, dosage of a therapeutic agent or their
combination.
57. A library of spectral signatures of field of view obtained from
the method of any one of claims 13, 26 and 42.
Description
FIELD OF INVENTION
[0002] This invention is directed to the use of in-vivo
hyperspectral imaging to monitor angiogenesis. Specifically, the
invention provides systems and methods of obtaining hyperspectral
images of a field of view comprising an area sought to be
monitored.
BACKGROUND OF THE INVENTION
[0003] Although traditional medical imaging modalities such as
computed tomography and magnetic resonance imaging have
significantly advanced our understanding of cancer behavior and
have provided a means for objectively determining tumor response to
anticancer therapeutics, historical definitions of tumor response
such as the WHO criteria, varied greatly among researchers,
creating a serious challenge in reporting results in a consistent
manner.
[0004] In an attempt to address these limitations, an international
collaboration of academicians, industry representatives, and
regulatory authorities established a unified, international
standard for tumor response assessment in 2000, termed the Response
Evaluation Criteria in Solid Tumors (RECIST), that was based on CT
and MRI linear measurements of lesion size. The National Cancer
Institute (NCI) has called for the improvement of the RECIST
methodology since these anatomical imaging techniques have proved
to be less than optimal predictors of therapeutic response in
comparison to the detection of molecular events that may better
correlate with diagnosis, staging, prognosis and response to cancer
therapy.
[0005] Classical hyperspectral imaging instruments are based on
changing filters using a liquid crystal tunable filter LCTF,
Acousto Optic Tunable filter (AOTF) or an interferometer. The
consequence is that these devices cannot be used if the objects in
the field of view (FOV) move or change during image/data
acquisition. Likewise, classical wavelength dispersive instruments
use diffraction gratings with poor light throughput and overlapping
spectral orders.
[0006] Therefore there is a need for an objective and reproducible,
accurate and sensitive method of monitoring and imaging the
response of organs, tissues and neoplastic growth to therapy, in
the life science community.
SUMMARY OF THE INVENTION
[0007] In one embodiment, the invention provides a hyperspectral
imaging system comprising: an electromagnetic energy source;
coupled to a prism and reflector imaging spectroscopy system
(PARISS) equipped with an imaging probe comprising a plurality of
customized bundle of structured optical fibers that can be uses as
a spatially resolved imaging probe.
[0008] In another embodiment, provided herein is a method of
acquiring in-vivo hyperspectral image from a subject, comprising:
selecting a field of view (FOV) of the subject; attaching a
plurality of customized, bundle of structured optical fibers that
can be uses as a spatially resolved imaging probe, wherein the
customized, spatially structured fiber-optic probes are operably
linked to a hyperspectral imaging system comprising: an
electromagnetic energy source; coupled to a prism and reflector
imaging spectroscopy system (PARISS); illuminating the field of
view using the hyperspectral imaging system; and collecting an
in-vivo hyperspectral image.
[0009] In one embodiment, the invention provides method of
monitoring neoplasia of a tissue in a subject, comprising the step
of obtaining a hyperspectral image of a field of view of an area
sought to be monitored; and comparing the image to a standard.
[0010] In another embodiment, the invention provides a method of
imaging a natural history or response to therapy of lesions of the
skin, oropharynx, esophagus, bladder, or intra-abdominal lesions
accessed through laparoscopy, comprising the step of obtaining a
hyperspectral image of a field of view of an area sought to be
monitored in the lesions of the skin, oropharynx, esophagus,
bladder, or intra-abdominal lesions accessed through laparoscopy;
and comparing the image to a standard.
[0011] In another embodiment, the invention provides a library of
spectral signatures of field of view obtained from the methods
described herein.
[0012] Other features and advantages of the present invention will
become apparent from the following detailed description examples
and figures. It should be understood, however, that the detailed
description and the specific examples while indicating preferred
embodiments of the invention are given by way of illustration only,
since various changes and modifications within the spirit and scope
of the invention will become apparent to those skilled in the art
from this detailed description.
BRIEF DESCRIPTION OF THE DRAWINGS
[0013] The invention will be better understood from a reading of
the following detailed description taken in conjunction with the
drawings in which like reference designators are used to designate
like elements, and in which:
[0014] FIG. 1 shows a photograph of the prototype in vivo
MACRO-PARISS hyperspectral imaging system. Arrows indicate the
fiber-optic imaging probe, an imaging spectrometer, a QICAM camera,
an observed image camera, and laptop that controls the cameras
(High intensity tungsten halogen lamp not shown);
[0015] FIG. 2 wherein the solid line shows the MIDL wavelength
calibration lamp spectrum after normalization, and the dotted line
before normalization. Note that following normalization, the
spectral range around 436 nm and 900 nm increased in intensity, and
the 546 nm line was almost unchanged;
[0016] FIG. 3 wherein the line with red diamond markers shows the
QE profile of the camera. The line with blue diamonds presents the
profile as rendered by the MACRO-PARISS system and is a convolution
of the QE of the camera, the actual spectral profile of the lamp,
coatings, fiber optic, and all other optical elements. The curve
with open circles is the profile in the NIST-certified profile; and
the superimposed curve with solid circles is the normalized
MACRO-PARISS profile of the lamp following normalization;
[0017] FIG. 4 shows a photograph of the 9 areas designated for
hyperspectral imaging. Note the clear juxtaposition of vascular and
non-vascular areas. This consistent vascular anatomy to allowed for
analysis of inter- and intra-animal imaging variability of
generated spectral data. Acquiring images of the same area on
different mice as well as imaging the same area on the same mouse
but on different days was possible;
[0018] FIG. 5 shows (A) radiometric Master Spectral Library (MSL).
Light colors tend towards areas in the FOV without vasculature and
those in redder colors indicate the presence of blood vessels. (B)
Using a MCC of 99%, spectral histograms depicting percent
composition of spectral objects were generated from Area 1
(vascular). The components of the histogram are a result of our
ability to display and characterize spatial heterogeneity. Images
were acquired on three separate days on the same mouse. The
spectral image to the far right represents the same spectral data
in spectral format. Note the consistency in spectral signatures in
histogram and spectral format, supporting the notion that this HSI
system can generate reproducible signatures over comparable regions
of interest. Spectral histograms for all nine areas were similarly
generated for analysis;
[0019] FIG. 6 shows spectral histograms of Area 1 (top
row-vascular) and Area 4 (bottom row-non-vascular). Note the
spectral differences in vascular regions of the skin when compared
to non-vascular regions. All spectral histograms were generated
from an MSL at 99% MCC;
[0020] FIG. 7 shows spectral histograms of Areas 5 and 9 closely
resembled a vascular signature. Note the "hybrid" spectral
histogram displayed in Areas 6 and 7. Spectral images are displayed
next to their respective histogram. Movement of the probe from a
vascular to a non-vascular area results in a significant and
step-wise change in spectral signatures;
[0021] FIG. 8 shows reversion of a vascular spectral signature
(Area 1) to a non-vascular spectral signature after the removal of
its blood supply. Area 4 did not change before and after the
removal of the blood supply. All spectral histograms are composites
of triplicate imaging acquisitions;
[0022] FIG. 9 shows an embodiment of spatial distribution of the
PARISS remote fiber optic probe. Different fibers collect light
from different areas of the sample, fibers that are mapped to the
slit can then be reconstructed to produce an image. Light source
(1), Lens focuses light source on illumination fibers (2), End of
fiber tip showing illumination fibers. Can be arrayed as a bundle
or a line depending on illumination source (3), Bifurcated fiber
integrates illumination and light collection fibers (4) End tip of
fiber that touches tissue (proximal end) (10). Shows illumination
fibers and collection fibers (5). Tip of collection fibers (distal
end) arrayed along a "line" (6) Lens to transfer image of
collection fibers onto the entrance slit (8) of the PARISS
spectrometer (9) (7). Entrance slit of PARISS spectrometer (9) (8).
PARISS spectrometer (9); and
[0023] FIG. 10 shows spectral analysis of an area of "Blood at 45
degrees tilt", MCC=99.5%. A. Proximal fiber tip, that touches
tissue, identifying the location of individual fibers arrayed at
the distal end and focused onto the slit of the spectrometer. B.
Array of fibers originating at the proximal end of the fiber
bundle. Each numbered fiber at the distal end can be traced to a
given location at the proximal end. C. Color codes corresponding to
"fingerprint" spectra in the MSL. D. A histogram showing the ratio
of MSL spectra that meet a given minimum correlation coefficient
(MCC). In this case the histogram is that of the acquisition shown
in (E). E. A spectral image taken through the PARISS spectrometer
of the distal end of the fiber bundle. Software assigns a color
code according to whether a given MMC. No color code is assigned if
the spectrum from a given fiber fails to meet a MCC target.
DETAILED DESCRIPTION OF THE INVENTION
[0024] This invention relates in one embodiment to the use of
in-vivo hyperspectral imaging to monitor angiogenesis. In another
embodiment, the invention relates systems and methods of obtaining
hyperspectral images of a field of view comprising an area sought
to be monitored.
[0025] In one embodiment, Hyperspectral imaging (HSI) refers to a
form of optical imaging that is used by the Remote Earth Sensing
Community for environmental studies, and is typically used in
radiometric mode. In another embodiment, HSI refers to any
instrument that faithfully digitizes an analog spectrum presented
by the field of view (FOV), requiring in another embodiment, that
each spectrum be characterized by a large number of data points. In
one embodiment, in the case of a spectrometer, a large number of
data points refers to greater than 600 wavelength data points over
the entire spectral range. This correlates with a spectral
resolution that will be better than 5 nm at wavelengths below 600
nm nad better than 10 nm at wavelengths up to 920 nm.
[0026] In one embodiment, the induction of new blood vessel
formation is a prominent feature of solid tumors and it is well
established that tumor size beyond 2 mm.sup.3 requires the
construction of a vessel network. Mean vascular density is
correlated in another embodiment, to progression of cutaneous
melanoma. In one embodiment, a correlation between overall survival
in melanoma patients and tumor microvessel density exists,
supporting the notion of a vascular density gradient and vertical
tumor progression.
[0027] In another embodiment, a HSI system can generate
quantifiable spectral differences in a reproducible manner between
vascular and non-vascular regions of skin. Accordingly and in one
embodiment, provided herein is a hyperspectral imaging system
comprising: an electromagnetic energy source; coupled to a prism
and reflector imaging spectroscopy system (PARISS) equipped with an
imaging probe comprising a plurality of customized bundle of
optical fibers that can be uses as probes. In one embodiment,
diffraction gratings diffract light into second and higher orders
with the consequence that efficiency degrades and longer
wavelengths can be polluted by commingled second order diffraction,
conversely, Prisms refract light into one order and present the
highest possible light transmission. In another embodiment, the
PARISS system used in the methods provided herein, is prism based
and presents ALL wavelengths simultaneously, therefore, in one
embodiment movement in the tissue imaged does not affect a spectral
characterization.
[0028] In one embodiment, the HSI system used in the methods and
compiling of the spectral images and libraries described
hereinbelow incorporates a high intensity tungsten halogen lamp
coupled with a PARISS (Prism and Reflector Imaging Spectroscopy
System)--and customized bundle of structured optical fibers that
can be uses as a spatially resolved imaging probe. The high
intensity tungsten lamp is used in another embodiment, as the
source of light because since it provides a greater amount of red
intensity, which in one embodiment, is ideal for penetrating and
probing living tissues. In another embodiment, since many variables
within an in vivo skin section such as pH, protein interactions,
temperature, and ionic concentration preclude a linear mixture of
spectral colors, the PARISS system used in the methods and
compiling of the spectral images and libraries described
hereinbelow is able to process non-linear high resolution spectral
data, thereby making it suitable for the evaluation sought.
[0029] In one embodiment, the light source is composed of a simple
light bulb, or a flash lamp, or another light source or combination
of sources, or it may be a complex form including gateable or
triggerable electronics, a light source, a filter element, a
transmission element such as an optical fiber, a guidance element
such as a reflective prism, and other elements intended to enhance
the optical coupling of the light from the emitter to the skin or
FOV under study in other embodiments. The light source may be
continuous, pulsed, or even analyzed as time, frequency, or
spatially resolved. The emitter may consist of a single or multiple
light emitting elements. In another embodiment, optical coupling
used to operably link the light source, or the imaging probe refers
to the arrangement of a light source (or light detector) in such a
way that light from the source (or detector) is transmitted to (or
detected from) the FOV, allowing passage through the tissue and
possible interaction with a contrast agent or in another
embodiment, a detectable molecular probe or marker. This may
require in one embodiment, the use of optical elements such as
lenses, filters, fused fiber expanders, collimators, concentrators,
collectors, optical fibers, prisms, mirrors, or mirrored surfaces
and their combination.
[0030] As would be appreciated by a person skilled in the art, the
use of LED light source of various colors is also encompassed by
the methods and systems described herein, to produce in certain
embodiments, off-white light with emphasis on certain ranges of the
spectrum, thereby obtaining a different response spectra. In
certain embodiment, the light source used in the methods and
compositions described herein, is optimized to the tissue being
imaged, or the pathology observed or monitored. It will be
appreciated that the spectra used to image the response of a
angiogenesis on the skin, may be different in one embodiment from
the optimal spectra used to obtain an intra-abdominal lesions
accessed through laparoscopy.
[0031] In one embodiment, spectral images are generated using a
charge-coupled device (CCD) detector. In another embodiment, the
plurality of customized bundle of structured optical fibers that
can be uses as a spatially resolved imaging probes used in the
systems provided herein are spatially discrete. In another
embodiment, the prism and reflector imaging spectroscopy system
comprises an imaging spectrometer integrated with a camera; and a
second camera, which in certain embodiments is QICAM camera, acting
as an observed image camera. In one embodiment, the plurality of
customized bundle of structured optical fibers that can be uses as
a spatially resolved imaging probes comprise illumination probes
and signal collection probes, wherein in yet another embodiment the
collection probes to are arrayed along a slit thereby collecting
images onto an entrance slit in the a prism and reflector imaging
spectroscopy system
[0032] In one embodiment, establishing use of the systems provided
herein in discriminating vascular and non-vascular areas allows for
the assessment of physiologic and pathologic processes in the tumor
microenvironment. Utilizing the systems provided herein coupled to
a bundle of structured optical fibers that can be uses as a
spatially resolved imaging probe, the reproducibility and
robustness of the spectral signatures derived from comparable
regions of interest is evident.
[0033] In one embodiment, performance optimization of the HSI
system in generating reproducible and unique spectral signatures is
determined by three factors. In one embodiment a high intensity
tungsten lamp, used as a light source, allows the emission of light
in the red range, which experiences less absorption by tissue. In
another embodiment, an appropriate acquisition parameters are set,
which increase the signal-to-noise ratio thereby reducing single
acquisition variability. In one embodiment, when creating the
master spectral library (MSL), only spectra that comes from the
center of the signal collection fibers is entered, leading to a
more robust MSL.
[0034] In one embodiment, the term "Signal to Noise ratio" refers
to the ratio of the strength of a target signal to the background
noise. This can be increased either by improving the target signal
in one embodiment, or by reducing the background noise or their
combination in other embodiments.
[0035] In one embodiment, the systems described herein, used in the
methods and compiling of the spectral images and libraries
described herein comprise a plurality of fiber optic probes,
consisting in one embodiment of a plurality of illumination fibers
and signal collecting fibers. In another embodiment, the
fiber-optic probe consists of 17 illumination fibers and 18 signal
collection fibers, which in one embodiment, are randomly
distributed, with each fiber having in another embodiment, a
50-micron core. Signal collection fibers are arrayed in one
embodiment, along a slit and imaged onto the PARISS entrance slit.
In one embodiment, each fiber delivers spectral information from a
different point in the FOV. In another embodiment, the in vivo
delivery of spatially discrete information differentiates the
system described herein from traditional spectral modalities that
acquire a single, homogenized, all encompassing, data-point. In one
embodiment, the number of illumination fibers and signal collection
fibers are optimized based on the area of the FOV sought to be
imaged, the core dimensions and the resolution sought. In one
embodiment, the fiber-optic probe consists of 15 illumination
fibers and 16 signal collection fibers. In another embodiment, the
fiber-optic probe consists of 20 illumination fibers and 21 signal
collection fibers. In another embodiment, the fiber-optic probe
consists of no less than 15 illumination fibers and no less than 15
signal collection fibers. (See e.g., FIG. 9)
[0036] In one embodiment, spatially structured fiber-optic probes
consisting of illumination fibers and signal collection fibers,
which in one embodiment, are randomly distributed, with each fiber
having in another embodiment, a core that is optimized for the
underlying application, the size of the FOV the location and other
factors in other discrete embodiment. As would be appreciated by a
person skilled in the art, the core diameter can be adjusted to
provide the optimal signal to noise ratio.
[0037] In one embodiment, the fiber-optic probe consists of 15
illumination fibers and 16 signal collection fibers may also be
partially attached to a cuff as well. The 15 illumination fibers
are cuffed together in one embodiment and placed on a tissue or
organ of the subject in another embodiment of the methods and
systems described herein. Likewise and in another embodiment, the
16 signal collection fibers may also be partially attached to a
cuff placed on the same or different body organ or tissue, such as
skin in another embodiment. While the cuff encircles the entire
number of fibers, it is within the scope of the invention (and
definition of the term `cuff`) to include a series of discrete
cuffs or patch-like elements distributed around the tissue or organ
of the subject, as points to which the illumination or collection
fibers are attached.
[0038] In one embodiment, the embodiments of the systems described
hereinabove and their inherent variables are used to carry out the
methods provided herein. Accordingly and in one embodiment,
provided herein is a method of acquiring in-vivo hyperspectral
image from a subject, comprising: selecting a field of view (FOV)
of the subject; attaching a plurality of customized remote
fiber-optic probes, wherein the customized remote fiber-optic
probes are operably linked to a hyperspectral imaging system
comprising: an electromagnetic energy source; coupled to a prism
and reflector imaging spectroscopy system (PARISS) equipped with an
imaging probe; illuminating the field of view using the
hyperspectral imaging system; and collecting an in-vivo
hyperspectral image.
[0039] In another embodiment, provided herein is a method of
acquiring in-vivo hyperspectral image from a subject, comprising:
selecting a field of view (FOV) of the subject; attaching a
plurality of customized remote fiber-optic probes, wherein the
customized remote fiber-optic probes are operably linked to a
hyperspectral imaging system comprising: an electromagnetic energy
source; coupled to a prism and reflector imaging spectroscopy
system (PARISS) equipped with an imaging probe; illuminating the
field of view using the hyperspectral imaging system; collecting an
in-vivo hyperspectral image; and compiling a unique spectral
signature of the field of view.
[0040] In one embodiment, each fiber in the bundle of the systems
probe presents light to the PARISS slit according to its respective
location in the FOV, providing in another embodiment spatial
information and allows to characterize field heterogeneity and
compile a spectral signature of the FOV. In one embodiment, the
spectral signature comprises the location and amplitude of signal
from every point in the FOV of the probe, thereby generating a
hyperspectral data cube consisting in one embodiment of wavelength,
or spatial and graphic information and their combination in another
embodiment (See e.g. FIG. 10).
[0041] In one embodiment, the spatial resolution of the fibers is
used for low resolution imaging. In another embodiment, the systems
and methods provided herein have a considerable advantage over
single, or pairs, of fibers.
[0042] In one embodiment, HSI systems used in the methods and
compiling of the spectral images and libraries described
hereinbelow comprising the fiber-optic attachment described
hereinabove in the form of a plurality of customized bundle of
structured optical fibers that can be uses as a spatially resolved
imaging probes comprise illumination probes and signal collection
probes, provides reproducible vascular or non-vascular signatures
and their combination in other embodiments. In one embodiment, as
the FOV moves from vascular to non-vascular areas, the acquired
spectra change in a step-wise predictable fashion, allowing the use
of the methods and compiling of the spectral images and libraries
described herein.
[0043] In one embodiment, the invention provides images obtained by
the embodiments of the methods described herein. In one embodiment,
provided herein is a method of monitoring angiogenesis in a subject
in response to a therapeutic treatment, comprising obtaining a
hyperspectral image of a field of view of the area sought to be
monitored; and monitoring changes in the spectral image in response
to the therapeutic treatment. In one embodiment, obtaining a
hyperspectral image of a field of view of the area sought to be
monitored comprises attaching a plurality of customized remote
fiber-optic probes to the area sought to be monitored, wherein the
customized remote fiber-optic probes are operably linked to a
hyperspectral imaging system comprising: an electromagnetic energy
source; coupled to a prism and reflector imaging spectroscopy
system (PARISS) equipped with an imaging probe; illuminating the
field of view using the hyperspectral imaging system; and
collecting an in-vivo hyperspectral image. In one embodiment, the
PARISS data processing algorithms is used for thresholding in
association with histograms. These can be set in another
embodiment, to "grade" the level of angiogenesis by providing
ratios of spectrally vascular to spectrally non-vascular
regions.
[0044] In another embodiment, provided herein is a method of
monitoring angiogenesis in a subject in response to a therapeutic
treatment, comprising obtaining a hyperspectral image of a field of
view of the area sought to be monitored; and monitoring changes in
the spectral image in response to the therapeutic treatment,
whereby obtaining a hyperspectral image of a field of view of the
area sought to be monitored comprises attaching a plurality of
customized remote fiber-optic probes to the area sought to be
monitored, wherein the customized remote fiber-optic probes are
operably linked to a hyperspectral imaging system comprising: an
electromagnetic energy source; coupled to a prism and reflector
imaging spectroscopy system (PARISS) equipped with an imaging
probe; illuminating the field of view using the hyperspectral
imaging system; collecting an in-vivo hyperspectral image; and
compiling a unique spectral signature of the field of view.
[0045] In one embodiment, the gradual spectral signature change,
which occurs in a predictable fashion based on the relevant
regional presence of vascularity, the HSI system described herein
provides a capability to monitor subtle changes in tumor
vascularity before and after therapeutic intervention. In one
embodiment, the angiogenesis sought to be monitored is associated
with cutaneous inflammatory and cancerous lesions. In another
embodiment, the angiogenesis sought to be monitored is associated
with gastrointestinal lesions via colonoscopy or esophagoscopy, and
during surgery such as in lymph node assessment. In another
embodiment, the angiogenesis sought to be monitored is associated
with differentiating oxy- and deoxyhemoglobin as a surrogate marker
of tumor hypoxia.
[0046] In certain embodiments, the use of spatially resolved fibers
or spatially structured fiber probes in other embodiments, or
imaging probe as described herein, enables the characterization of
tissue in which it is expected that at a micro level some areas
will be vascular and some not. The ratio of vascular to non
vascular areas, as resolved using the methods described herein, at
a micro level, assists in another embodiment, in determining the
spatial borders of a tumor, which in another embodiment, may be
used in automated cancer detection.
[0047] In one embodiment, provided herein is a method of monitoring
angiogenesis in a subject in response to a therapeutic treatment,
comprising obtaining a hyperspectral image of a field of view of
the area sought to be monitored; and monitoring changes in the
spectral image in response to the therapeutic treatment, wherein
obtaining a hyperspectral image of a field of view of the area
sought to be monitored comprises attaching a plurality of
customized remote fiber-optic probes to the area sought to be
monitored, wherein the customized remote fiber-optic probes are
operably linked to a hyperspectral imaging system comprising: an
electromagnetic energy source; coupled to a prism and reflector
imaging spectroscopy system (PARISS) equipped with an imaging
probe; illuminating the field of view using the hyperspectral
imaging system; collecting an in-vivo hyperspectral image;
compiling a unique spectral signature of the field of view; and
comparing the spectral signature of the field of view with the
spectral signature of the same field of view obtained from a
subject exhibiting angiogenesis in one embodiment, or a subject not
exhibiting angiogenesis in another embodiment.
[0048] In one embodiment the methods of data acquisition provided
herein, may be useful to the clinic as a reliable adjunct to the
pathologist, oncologist, surgeon and dermatologist in monitoring
tumor response after therapeutic intervention.
[0049] In one embodiment, each fiber in the bundle or cuff in
another embodiment, presents light to the PARISS slit according to
its location in the FOV. Spectral and illumination variations from
fiber to fiber are dependant on; localized variations in the FOV
(spatial inhomogeneity) in one embodiment, or variations in tilt,
variations in pressure, individually damaged fibers, motion or
their combination in other embodiments. Accordingly, in certain
embodiments the spectral image obtained using the systems described
herein in the methods provided herein is adjusted or normalized to
the factors described herein, to be compared to a standard spectral
image library. In one embodiment once images obtained using the
systems described herein in the methods provided herein are
normalized, they are incorporated into the libraries provided
herein.
[0050] When the target tissue is dry in one embodiment, an index
matching fluid is used such as oil in one embodiment, or glycerin
in another, to better couple the proximal end of the fibers with
the tissue under examination. In another embodiment this increases
light transmission and reduce the effects of wear on the ends of
the fibers". In another embodiment, an index matching material is
disposed between the dry tissue and the proximal end of the fibers,
for maintaining a constant and matched index for the light directed
into the tissue and the light reflected from the tissue. In one
embodiment, an index matching gel reduces large index of refraction
changes that would occur normally between a dry tissue and a gap of
air. These large changes result in Fresnel losses that are
especially significant in a reflectance based analysis, which
creates significant changes in the spectral signal. According to
one embodiment of the present invention, the indexing matching
material is a chloro-fluoro-carbon gel. This type of indexing
material has several favorable properties. First, the
chloro-fluoro-carbon gel minimally impacts the spectral signal
directed through the gel. Second, this indexing matching material
has a high fluid temperature point so that it remains in a gel-like
state during the analysis and under test conditions. Third, this
gel exhibits hydrophobic properties so that it seals the sweat
glands so that sweat does not fog-up (i.e., form a liquid vapor on)
proximal end of the fiber (tip). And fourth, this type of index
matching material will not be absorbed into the stratum corium of
skin during the analysis.
[0051] In one embodiment, provided herein is a library of spectral
signatures of field of view obtained from a subject undergoing
angiogenesis therapy, wherein the field of view comprises a tumor
area, a vascular area, a non-vascular area, a cutaneous
inflammatory lesion, a cancerous lesion, a gastrointestinal lesion
or a combination thereof.
[0052] In one embodiment, the method described herein, are capable
of being applied to other clinical applications such as early
detection of neoplasia assessing remission or recurrence, or in
other embodiment, monitoring either or both the natural history or
response to therapy of lesions of the skin in one embodiment, or
oropharynx, esophagus, bladder or potentially intra-abdominal
lesions accessed through laparoscopy in other discrete embodiments.
In another embodiment, the methods of obtaining a hyperspectral
images using the systems described herein, can collect spectral
signatures that are attributed to angiogenesis in one embodiment,
or other alterations in a field of view, including metabolic
changes, necrosis, inflammation, or neoplastic transformation in
other discrete embodiments of the methods described herein. The
methods and systems described herein, are used in another
embodiment to provide evidence of therapeutic response to cancer
therapy (chemotherapy or various forms of radiation) or other
therapies such as photodynamic therapy in certain embodiments,
reflected as changes in spectral characteristics due to altered
angiogenesis or other metabolic changes or necrosis as
described.
[0053] In one embodiment, provided herein is a method of monitoring
neoplasia of a tissue in a subject, comprising the step of
obtaining a hyperspectral image of a field of view of an area
sought to be monitored; and comparing the image to a standard,
whereby the step of obtaining a hyperspectral image of a field of
view of the area sought to be monitored comprises attaching a
plurality of customized remote fiber-optic probes to the area
sought to be monitored, wherein the customized remote fiber-optic
probes are operably linked to a hyperspectral imaging system
comprising: an electromagnetic energy source; coupled to a prism
and reflector imaging spectroscopy system (PARISS) equipped with an
imaging probe; illuminating the field of view using the
hyperspectral imaging system; and collecting an in-vivo
hyperspectral image.
[0054] In another embodiment, provided herein is a method of
imaging a natural history or response to therapy of lesions of the
skin, oropharynx, esophagus, bladder, or intra-abdominal lesions
accessed through laparoscopy, comprising the step of obtaining a
hyperspectral image of a field of view of an area sought to be
monitored in the lesions of the skin, oropharynx, esophagus,
bladder, or intra-abdominal lesions accessed through laparoscopy;
and comparing the image to a standard, whereby the step of
obtaining a hyperspectral image of a field of view of the area
sought to be monitored comprises attaching a plurality of
customized remote fiber-optic probes to the area sought to be
monitored, wherein the customized remote fiber-optic probes are
operably linked to a hyperspectral imaging system comprising: an
electromagnetic energy source; coupled to a prism and reflector
imaging spectroscopy system (PARISS) equipped with bundle of
structured optical fibers that can be uses as a spatially resolved
imaging probe; illuminating the field of view using the
hyperspectral imaging system; and collecting an in-vivo
hyperspectral image.
[0055] In one embodiment, the compiled unique hyperspectral
signature images are used to make a master spectral library. The
term "library of hyperspectral images" refers in one embodiment to
the collection of hyperspectral data that is being generated
employing the systems and methods disclosed herein. The basis of
choice of the library of hyperspectral images used in the methods
described herein, will vary in another embodiment, with the
application.
[0056] In one embodiment, the library provided herein is used for
identifying the type of neoplastic process, or in other embodiments
the degree, responsiveness to treatment, pathology, and the like in
other embodiments. The fact that complete spectra are available,
allows in one embodiment, to identify the different responses. The
samples upon which the library is built may be categorized, or
otherwise diagnosed, as diseased or non-diseased by a variety of
methods. In one embodiment, a pathologist utilizes conventional
procedures to make such a determination. In another embodiment, the
diagnosis is made by conventional histological techniques,
including conventional histochemical and/or biochemical techniques.
In certain embodiment, following such diagnosis, the spectral image
of the sample is obtained using the systems and methods described
herein. Accordingly, and in one embodiment, the database, or
library, may include a digital spectrum library, and/or a library
of desired spectral features as described herein, stored in a
computer. In other forms of the invention, the spectroscopic data
may be compared graphically or by other similar methods. If
desired, a background adjustment may be made by having the software
subtract from the spectra analyzed the background reflectance
spectra of normal tissue, both with and without stain, including a
baseline spectrum of the patient's normal tissue.
[0057] The term "about" as used herein means in quantitative terms
plus or minus 5%, or in another embodiment plus or minus 10%, or in
another embodiment plus or minus 15%, or in another embodiment plus
or minus 20%.
[0058] The term "subject" refers in one embodiment to a mammal
including a human in need of therapy for, or susceptible to, a
condition or its sequelae. The subject may include dogs, cats,
pigs, cows, sheep, goats, horses, rats, and mice and humans. The
term "subject" does not exclude an individual that is normal in all
respects.
[0059] The following examples are presented in order to more fully
illustrate the preferred embodiments of the invention. They should
in no way be construed, however, as limiting the broad scope of the
invention.
EXAMPLES
Materials and Methods
Hyperspectral Imaging System Instrumentation
[0060] The in vivo hyperspectral imaging system consists of a
spatially discrete, multi-fiber optic imaging probe (LightForm,
Inc., Hillsborough, N.J.), an imaging spectrometer integrated with
a QICAM camera (Q Imaging, Burnaby Canada), and a second QICAM
acting as an observed image camera (FIG. 1).
[0061] The fiber-optic probe consists of 17 illumination fibers and
18 signal collection fibers, randomly distributed, with each fiber
having a 50-micron core. Signal collection fibers were arrayed
along a slit and imaged onto the PARISS entrance slit. Each fiber
delivered spectral information from a different point in the
FOV.
[0062] The MACRO-PARISS system (LightForm, Inc., Hillsborough,
N.J.) is a prism based imaging spectrometer that originated within
the remote Earth sensing community. This system was chosen because
of its very high light transmission (>90%) characteristics
typical of prism systems. The imaging spectrometer portion operates
in spectrograph configuration in which all wavelengths between 365
and 920 nm are presented simultaneously. Acquiring all wavelengths
within a single fast acquisition accommodates movement in the FOV
without affecting the integrity of a spectral acquisition. In this
study, a range of 450 to 920 nm was chosen.
Calibration, Normalization and Validation
[0063] The MACRO-PARISS spectrometer was first wavelength
calibrated using a multi-ion discharge lamp (MIDL), (LightForm,
Inc., Hillsborough, N.J.) that emits Hg+, Ar+ and inorganic
fluorophores (FIG. 2). Each pixel in a column of the spectrum CCD
corresponds to a specific wavelength. The MIDL lamp provides the
absolute wavelength information provided by the ion emission lines.
The MACRO-PARISS software provides an algorithm that matches the
spectral features to pixel with subsequent calibration of the
entire spectrometer. The wavelength accuracy was validated to be
better than 0.5 nm over the entire spectral range.
[0064] Spectral resolution was confirmed by measuring the number of
pixels that covered the full width at half maximum (FWHM) at the
436 nm Hg line. The wavelength spread at FWHM was determined to be
1.2 nm+/-0.25 nm.
[0065] The MACRO-PARISS spectrometer was then normalized to remove
instrumental contributions due to the Quantum Efficiency (QE) of
the camera, coatings on lenses, the fiber-optic and prism
transmission properties, and reflectivity of mirrors in the system.
A NIST-certified halogen lamp (NCHL) (Model LS-1-Cal, Cert #1013,
Ocean Optics, Inc, Dunedin Fla.) was used that was supplied with a
spectral profile in ASCII listing wavelength versus power
(.mu.W/cm.sup.2/nm). This is the profile any radiometric
spectrometer would report when characterizing this lamp. The
MACRO-PARISS software incorporates the algorithm to enable this
normalization. FIGS. 2 and 3 show the profile of the NCHL reported
by the MACRO-PARISS spectrometer before and after correction, and
the QE of the camera. The corrected profile is seen to be a perfect
match to the profile shown on the lamp's certificate.
Light Source, Calibration and Methods
[0066] A high intensity tungsten halogen lamp set to the highest
brightness was used as the light source for the fiber-optic probe
to increase signal-to-noise data at the ends of the spectral range
and for greater red intensity. The wavelength range from 450 to 920
nm, especially in the near IR, above 650 nm, experiences less
absorption in living tissues and would theoretically provide
valuable tissue spectral information otherwise not accounted for
with other light sources.
[0067] The entire spectral system, including the probe, was
wavelength-calibrated daily prior to imaging experiments to ensure
accuracy using the MIDL emission source as described above.
Spectral sensitivity was checked in the advanced display mode with
amplitude plotted against wavelength at the time of calibration.
After wavelength calibration, the system was normalized
radiometrically through the fiber-optic probe with the NIST
certified, halogen light source. A correction curve was then
generated in the MACRO-PARISS software to correct all acquisitions
to the standard.
[0068] Since each fiber in the bundle presents light to the
MACRO-PARISS slit according to its respective location in the FOV,
an attempt was made to account for a heterogeneous FOV by using a
fiber-optic probe with a relatively higher number of individual
fibers--as described in the instrumentation section. This would
provide spatial information and allow the characterization of field
heterogeneity. The fact that spatial information is delivered in
vivo differentiates this system from traditional spectral
modalities that acquire a single, homogenized, all encompassing,
data-point.
[0069] Additionally, the ability to map the 18 collection fibers to
the slit in order to generate a low resolution graphic image is
recognized and being addressed. Determining the location and
amplitude of signal from every point in the FOV of the probe,
becomes possible, generating a hyperspectral data cube consisting
of wavelength, spatial and graphic information.
In Vivo Spectral Acquisition and Acquisition Parameters
[0070] The branching artery of the SCID mouse ear was used as the
model because it has adjacent regions of vascular and non-vascular
skin. Nude mice were selected due to their ears having a consistent
vascular anatomy, allowing for a clear comparative analysis between
vascular and non-vascular acquisitions. Additionally, it was
anticipated that hair would contribute to a spectral signature and
hence, nude mice would circumvent this issue for these
examples.
[0071] Using an Institutional Animal Care and Use Committee
(IACUC)-approved protocol, three nude mice were anesthetized with a
ketamine/xylazine solution (125 mg/kg and 15 mg/kg, respectively)
via intraperitoneal injection. Subsequently, the probe was gently
placed with uniform pressure on the skin, ninety degrees to the
selected region for spectral acquisition.
Parameters to Reduce Single Acquisition Variability
[0072] Specific acquisition parameters were determined based on an
inverse relationship between movement variability of the user and
acquisition variability. The fact that all wavelengths were
acquired simultaneously ensured that there was no variability from
wavelength to wavelength; however, movement of the probe due to the
user could result in differences in the illuminated area. The
greater the acquisition times the lesser the acquisition
variability but the greater the chance of movement by the user.
[0073] An attempt was made to address the problem of spectral
sampling variability by using standard analytical sampling
procedures. Each in vivo skin acquisition consisted of 5 co-added
25 millisecond acquisitions 5-times; resulting in an improvement in
the signal-to-noise ratio (S/N) of a factor of 2.2 (square root of
5). Each area was sampled 5 times, which was found to be an optimal
balance where any investigator could hold still, without
compromising the acquisition quality and hence, maintain a
well-defined S/N ratio.
[0074] A total of nine distinct skin regions (Areas 1-9) were
imaged in triplicate on different days and subsequently compared
(FIG. 4). At the time of sacrifice, vascular and non-vascular skin
areas were imaged before and after stripping the vascular blood
supply to the ear via a 1 cm surgical incision made at the base of
the ear, with subsequent drainage of the blood to gravity.
Image Analysis Software
[0075] It was assumed that many physiologic variables within an in
vivo skin region would preclude a linear mixture of spectral
signatures. The MACRO-PARISS system accounts for this physiological
challenge since it is able to process non-linear spectral data from
spectral objects localized near each other. Master spectral
libraries (MSL) were generated from acquired spectral data through
a supervised classification using a spectral waveform cross
correlation analysis (SWCCA) algorithm. All library spectra were
saved in radiometric format (FIG. 5A). Spectra corresponding to
more vascular areas were pseudo-colored in redder colors and
non-vascular areas in bluer colors. Note that a minimum was formed
around 550 nm that probably corresponds to absorption by red blood
cells.
[0076] While previous studies have used a combination of automated
and supervised classification schema for generating MSL, a more
conservative strategy was taken, since an automated classification
would have included spectral data from the peripheral regions of
the fiber-optic collection fibers, leading to less robust spectral
signatures. Each spectrum that was manually added to the MSL was
designated a pseudo color. All spectra from the original FOV were
then correlated with the MSL, set to a minimum correlation
coefficient (MCC) of 99%. Only when the spectrum from the FOV
correlated at 99% or greater with an MSL spectrum, did the
pseudo-color replace the gray scale pixel. Additionally, histograms
of MSL pseudo-color members were generated to present spectral
ratios within the FOV and were used to compare vascular and
non-vascular skin regions.
Example 1
Reproducibility and Robustness of Acquired MACRO Hyper-Spectral
Signatures
[0077] The approach to testing the reproducibility of the MACRO-HSI
system consisted of imaging the vascular anatomy of the ear, since
it allowed to image the same designated areas on the same mouse on
different days, as well between mice. This approach would allow to
assess intra- and inter-animal imaging variability. All 4 areas of
the ear were imaged in triplicate on different days.
[0078] The Master Spectral Library (MSL) was used to generate all
ear spectral histograms. All spectra from the original FOV were
correlated with the MSL, set to a minimum correlation coefficient
(MCC) of 99%. Spectra from the spectral image from light captured
exclusively from the center (core) of the 18 collection fibers were
manually added. This served to create a more robust spectral
library since it did not include spectral data from the walls of
the collection fibers. Each fiber delivered a spectral signature
from a different point on the FOV.
[0079] FIG. 5B clearly demonstrates the consistency in spectral
histograms generated from the nine areas of variable vascularity
within the mouse ear.
Example 2
In Vivo Hyperspectral Differentiation of Vascular and Non-Vascular
Regions of the SCID Mouse Ear
[0080] Spectral histograms of vascular and non-vascular regions of
the skin were compared. FIG. 6 illustrates the spectral differences
in vascular regions (Areas 1 and 3) of the skin as compared to
non-vascular regions (Areas 2 and 4). The results indicate that
similar regions of interest with respect to vascularity
consistently generated a unique spectral signature. Further
characterization of the uniqueness of the vascular spectral
signature was made by imaging additional regions of the ear, Areas
5-9. These additional areas were chosen for analysis because they
all included visible vascularity but differed with respect to the
level of gravity pull and their distance from the base vascular
supply. It was sought to image a gradation of vascular prominence
as a potential surrogate marker for vessel formation and
retraction.
[0081] With these additional areas, it was hypothesized that Areas
5 and 9 would generate spectral signatures similar to 1 and 3
(i.e., a vascular signature) while Area 8 would closely resemble
Areas 2 and 4 (i.e., a non-vascular signature). Areas 6 and 7 might
behave in one of two ways: either 1) these areas would capture a
signature that would be a combination of vascular and nonvascular
signatures or 2) capture a vascular signature.
[0082] The results confirmed the hypothesis that the spectral
histograms of Areas 5 and 9 closely resembled Areas 1 and 3.
Additionally, it was found that Areas 6 and 7 displayed a "hybrid"
spectral histogram between that of Areas 5, 9 and Area 8 (FIG. 7).
This data led to the characterization of a unique vascular
signature.
[0083] Overall, these findings led to the conclusion that the SCID
mouse ear provides a model to establish a vascular hyperspectral
signature spectrum as one moves away from the base of the ear,
against gravity, towards to apex of the ear. This data also
supports the conclusion that the MACRO-PARISS HSI system
fiber-optic attachment provides reproducible vascular and
non-vascular signatures. As the FOV moved from vascular to
non-vascular areas, the acquired spectra changed in a step-wise
predictable fashion.
Example 3
Removal the Blood Supply from a Vascular Field with Subsequent
Spectral Acquisitions
[0084] At the time of mouse sacrifice, vascular and non-vascular
skin regions were imaged before and after stripping the vascular
blood supply to the ear via a 1 cm surgical incision made at the
base of the ear with subsequent drainage of the blood with gravity.
It was hypothesized that removal of the blood supply would lead to
vascular regions taking on a new non-vascular spectral signature.
Additionally, it was expected that no change in spectral signatures
in regions that were non-vascular prior to stripping the vascular
blood supply. FIG. 8 illustrates vascular regions taking on
non-vascular spectral signatures after the removal of their blood
supply. This approach further supports the conclusion that the
spectral signatures are indeed vascular and non-vascular.
[0085] Having described preferred embodiments of the invention with
reference to the accompanying drawings, it is to be understood that
the invention is not limited to the precise embodiments, and that
various changes and modifications may be effected therein by those
skilled in the art without departing from the scope or spirit of
the invention as defined in the appended claims.
* * * * *